[R] fixed and random effects question

[Prof Brian Ripley]

Consider the case of a response vs time for multiple measurements on 
subjects.  If one has both fixed and random effect for slope and 
intercept, it means there is a separate linear relationship with time for 
each subject (the random effect) and that the population average may have 
a non-zero slope and intercept (the fixed effect).

On Sat, 10 Apr 2004, Scott Rifkin wrote:

> On page 146 of Pinheiro & Bates, they mention models  where every fixed 
> effect has an associated random effect. What does it mean for a fixed 
> effect to have an associated random effect?

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595